It is well known that oxide scales formed on the steam side of the superheater and reheater tubes in power generation boilers sometimes spall after a long time operation. Spalling of oxide scales may cause blockage of the tubes as well as erosion of the first stage nozzles and buckets of the turbines. This study was carried out to obtain the knowledge about the spalling behavior of the steam oxide scale in stainless-steel tubes. SUS 321 HTB tube specimens were heated in steam atmosphere at 600-750°C for more than 300h, and then cooled with 50-170°Ch-1. Mass change and cross section of specimens were examined to evaluate the scale spalling. Obtained results are summarized as follows: (1) Micro-cracks were observed along the interface between the outer and inner oxide scales for all specimens but macroscopic spalling of the outer scale occurred only for the specimens heated at>700°C. (2) Extension of the interface separation was independent of heating temperature and cycles, and increased with inner scale thickness. (3) Spalling of outer oxide scales was strongly affected by heating temperature and heating cycles.
Four types of corrosion tests were carried out to determine a relationship corrosion rate, CR (Fe), carbon steel and ACM sensor output. Test 1-1 and 1-2 were laboratory tests under controlled relative humidity, RH, and amount of deposited sea salt, Ws Test 2-1 and 2-2 were outdoor exposed tests in Shimizu city, Shizuoka prefecture, and Nishihara town, Okinawa prefecture, respectively under conditions sheltered from rainfall. Corrosion rate of carbon steel coupons was determined by conventional mass loss method in the tests except in Test 1-2 for lower corrosion rate, to which QCM (Quartz Crystal Microbalance) technique was applied. The output of the ACM sensor, galvanic current, I, between carbon steel and silver, was recorded every ten minutes with temperature and relative humidity, and analyzed to be related to the corrosion rate, CR (Fe), as a formula: log CR(Fe)[mm/y]=0.378logQ[C/day]-0.636 (1) Where Q is daily average electricity, integral of ACM output I over a day. The ACM sensor can predict corrosion rates as low as 2μm/y through the equation (1), being slightly short of 0.2μm/y of QCM technique.
Corrosion resistance of the thermal sprayed coatings of Zn, Al, and Zn-15Al alloy exposed in the atmosphere for 22 years at Tokyo and Sapporo were evaluated by the electrochemical anodic polarization measurements, SEM observation and EPMA, comparing with the as-sprayed specimen. These coatings showed excellent resistance against the atmospheric corrosion for over 20 years. Among the 3 kinds of metals, Zn was the most superior comparing with that of Zn-Al alloy and Al. Corrosivity of the atmosphere caused by S and Cl was larger at Sapporo than at Tokyo. The fact suggests that the influence of the airborne salinity and the air pollution substances is more severe at the coast of Japan Sea rather than the coast of Pacific Ocean.
The electrochemical impedance for infinite and finite diffusion processes was described, and the locus of the impedance was discussed through comparison of the experimental results with the theoretical ones. The charge transfer resistance Rct and the polarization resistance Rp were compared with the current i and the slope of the polarization curve for the reduction of K3Fe(CN)6. The physical meanings of Rct and Rp in the diffusion impedance were discussed. It was confirmed that Rct-1 was related to the reaction rate. The theory to apply the diffusion impedance to the analysis of corrosion reaction was revealed.
Non-destructive evaluation of a component material by an acoustic impact method was fundamentally investigated for the simplified square samples with artificial defects. Frequency spectrum, which was obtained by a fast Fourier transform of an acoustic wave form, was discussed for the evaluation of the defects on the specimen. The analyses on the frequency spectrum recognized differences in frequency and amplitude between defective and non-defective materials. The mode analyses obtained by using laser displacement rigs suggested that many peaks in the frequency spectrum were appeared by the generation of multiple vibration modes in the square specimen. It was found that the existence of small defects did not change the vibration modes but shifted the location of node and loop on one of the vibration modes against the non-defect specimen. The common frequency both of the vibration and of the sound indicated the entirely same mode number. The intensities of the two detected waves at the same location on the specimen were, however, reversed between the vibration and the sound.